Providing an Instance Availability Estimate

1. A computer-implemented method for providing an instance availability estimate in a distributed execution environment, the method comprising performing computer-implemented operations for: collecting data defining actual boot times for one or more virtual machine instances, the one or more virtual machine instances having a plurality of combinations of instance types and operating systems; utilizing the collected data to compute an estimate of a time required to boot a virtual machine instance having each combination of instance type and operating system; storing the computed estimates in a historical boot times data store; receiving a request to launch a new virtual machine instance; in response to receiving the request to launch the new virtual machine instance, identifying an instance type and operating system for the new virtual machine instance; utilizing the instance type and operating system identified for the new virtual machine instance to retrieve an estimate of the time required to boot the new virtual machine instance from the historical boot times data store; utilizing the estimate of the time required to boot the new virtual machine instance to compute an instance availability estimate for the new virtual machine instance; and providing the instance availability estimate in response to the request to launch the new virtual machine instance.

2. The computer-implemented method of claim 1 , wherein computing an instance availability estimate for the new virtual machine instance comprises adding the estimate of the time required to boot the new virtual machine instance to an estimate of a time to begin booting the virtual machine instance in the distributed execution environment following receiving the request to launch the new virtual machine instance.

3. The computer-implemented method of claim 2 , wherein collecting the data defining the actual boot times for one or more virtual machine instances comprises collecting data defining the actual boot times for one or more virtual instances launched by customers of the distributed execution environment.

4. The computer-implemented method of claim 2 , wherein collecting the data defining the actual boot times for one or more virtual machine instances comprises collecting data defining the actual boot times for one or more virtual instances launched by a test harness service executing in the distributed execution environment.

5. The computer-implemented method of claim 4 , wherein the virtual machine instances launched by the test harness service comprise one or more combinations of instance types and operating systems for which insufficient data defining the actual boot times has been collected in order to generate the estimate of the time required to boot the virtual machine instances.

6. The computer-implemented method of claim 5 , wherein the estimate of the time required to boot a virtual machine instance having a combination of instance type and operating system comprises a 95% confidence estimate of the time required to boot a virtual machine instance having the combination of instance type and operating system.

7. A system for providing an instance availability estimate, the system comprising one or more processors configured to: execute, on at least one of the one or more processors, a data collection service configured to collect data defining actual boot times for one or more virtual machine instances, the one or more virtual machine instances having a plurality of combinations of instance types and operating systems; execute, on at least one of the one or more processors, a statistics computation service configured to utilize the data collected by the data collection service to compute an estimate of time required to boot a virtual machine instance having each combination of instance type and operating system and to store the computed estimates in a historical boot times data store; and execute, on at least one of the one or more processors, an instance launch manager configured to receive a request to instantiate a new virtual machine instance, in response to receiving the request, to identify an instance type and operating system for the new virtual machine instance, to utilize the instance type and operating system identified for the new virtual machine instance to retrieve an estimate of the time required to boot the new virtual machine instance from the historical boot times data store, to utilize the estimate of the time required to boot the new virtual machine instance to compute an instance availability estimate for the new virtual machine instance, and to provide the instance availability estimate in response to the request to instantiate the new virtual machine instance.

8. The system of claim 7 , wherein the estimate of the time required to boot the new virtual machine instance is generated by adding the estimate of the time required to boot the new virtual machine instance to an estimate of a time to begin instantiating the new virtual machine instance following receiving the request to instantiate the new virtual machine instance.

9. The system of claim 8 , wherein the data defining the actual boot times for one or more virtual machine instances is collected for virtual machine instances instantiated by customers of a distributed execution environment in which the virtual machine instances are instantiated.

10. The system of claim 8 , wherein the data defining the actual boot times for one or more virtual machine instances is collected for virtual machine instances instantiated by a test harness service executing in a distributed execution environment in which the virtual machine instances are instantiated.

11. The system of claim 10 , wherein the estimate of the time required to boot the virtual machine instance comprises a 95% confidence estimate of the time required to boot the virtual machine instance.

12. A non-transitory computer-readable storage medium having computer-executable instructions stored thereupon which, when executed by a computer, cause the computer to: collect data defining actual boot times for one or more virtual machine instances, the one or more virtual machine instances having a plurality of combinations of instance types and operating systems; utilize the collected data to compute an estimate of time required to boot a virtual machine instance having each combination of instance type and operating system; store the computed estimates in a historical boot times data store; receive a request to launch a new virtual machine instance; in response to receiving the request to launch the new virtual machine instance, identify an instance type and operating system for the new virtual machine instance; utilize the instance type and operating system identified for the new virtual machine instance to retrieve an estimate of the time required to boot the new virtual machine instance from the historical boot times data store; utilize the estimate of the time required to boot the new virtual machine instance to compute an instance availability estimate for the new virtual machine instance; and provide the instance availability estimate in response to the request to launch the new virtual machine instance.

13. The non-transitory computer-readable storage medium of claim 12 , wherein the one or more virtual machine instances are provided by a distributed execution environment.

14. The non-transitory computer-readable storage medium of claim 13 , wherein the instance availability estimate is further provided in response to receiving a request for the instance availability estimate at an application programming interface (API).

15. The non-transitory computer-readable storage medium of claim 13 , having further computer-executable instructions stored thereupon which, when executed by a computer, cause the computer to: store data defining the hardware upon which the one or more virtual machine instances were executed; and utilize the stored data defining the hardware to provide the instance availability estimate for a specific configuration of a virtual machine instance executing on the hardware.

16. The non-transitory computer-readable storage medium of claim 13 , having further computer-executable instructions stored thereupon which, when executed by a computer, cause the computer to: store data defining a region in which the one or more virtual machine instances were executed; and utilize the stored data defining the region in which instances were executed to provide the instance availability estimate for a specific configuration of a virtual machine instance executing in a region.

17. The non-transitory computer-readable storage medium of claim 13 , having further computer-executable instructions stored thereupon which, when executed by a computer, cause the computer to: store data defining the time at which the one or more virtual machine instance were executed; and utilize the stored data defining the time at which the one or more virtual machine instances were executed to provide the instance availability estimate for a specific configuration of a virtual machine instance that begins executing at a particular time.

18. The non-transitory computer-readable storage medium of claim 12 , wherein computing an instance availability estimate for the new virtual machine instance comprises adding the estimate of the time required to boot the new virtual machine instance to an estimate of a time to begin booting the virtual machine instance in the distributed execution environment following receiving the request to launch the new virtual machine instance.

19. The non-transitory computer-readable storage medium of claim 13 , wherein collecting the data defining the actual boot times for one or more virtual machine instances comprises collecting data defining the actual boot times for one or more virtual instances launched by customers of the distributed execution environment.

20. The non-transitory computer-readable storage medium of claim 13 , wherein collecting the data defining the actual boot times for one or more virtual machine instances comprises collecting data defining the actual boot times for one or more virtual instances launched by a test harness service executing in the distributed execution environment.